Infinity Loss / Inf Weights
Infinity loss or weights occur when loss values overflow to infinity, often from numerical instability in mixed-precision training.
Infinity loss or weights occur when loss values overflow to infinity, often from numerical instability in mixed-precision training.
What this failure is
Infinity Loss / Inf Weights is a Training Stability failure seen during ML training runs. Infinity loss or weights occur when loss values overflow to infinity, often from numerical instability in mixed-precision training. Common tags: Inf, Infinity, Loss, Numerical Stability.
Is this what broke your run? Paste your log.
You're reading about Infinity Loss / Inf Weights. Paste your own crash log or traceback below and get the real root cause for YOUR run, not this generic entry. No account, no card. Logs are masked at ingress and never saved to account history.
Want 14 days on the Scale plan?
Request an evaluation code. A verified workplace organization activates up to 50 diagnoses a day, alerts, history, and follow-up questions. No credit card or automatic subscription.
Why it happens (the mechanism)
Overflow in loss computation (e.g., log(0) in cross-entropy). Mixed precision with incorrect loss scaling. Division by very small number producing inf. Taken together, these mechanisms explain why the failure is reproducible, why it tends to surface on specific workloads or scales, and why generic mitigation attempts often fall short without addressing the underlying cause.
What you'll observe
- Loss becomes inf instead of NaN
- Weights contain inf values
- Training continues but model is corrupted
Common symptoms and what they mean
| Symptom | Why it happens |
|---|---|
| loss: inf in training logs | Overflow in loss computation (e.g., log(0) in cross-entropy) |
| Parameter contains inf values | Mixed precision with incorrect loss scaling |
| Gradient norm reports inf | Division by very small number producing inf |
Which systems are affected
- Mixed-precision training with large learning rates
- Models with log-softmax output
- Training with extreme loss values
How to confirm this is the problem
Use this checklist to test the hypothesis against a small reproduction. No single line proves the root cause, so preserve the preceding events and compare one variable at a time.
- ✓Reproduce the failure from a clean checkpoint/seed: the symptom must appear without warm-up state from a previous run.
- ✓Verified signal present: loss: inf in training logs
- ✓Verified signal present: Parameter contains inf values
- ✓Verified signal present: Gradient norm reports inf
- ✓A targeted fix from the "How to fix it" section eliminates or substantially reduces the symptom within one validation pass.
The fix and the prevention pattern
The root cause is on this page and stays free. A free account adds the exact remediation steps, saved history, and the fix on every entry in the encyclopedia.
Sign up free. Unlock the full analysisNo credit card. Daily allowance follows verified trust tier. Instant access.
Diagnose this failure in VS Code
Select the traceback or open the failed terminal, then run Denpex locally to see the initiating rank, collateral failures, exact fix, and verification command without uploading the log.
Install the free VS Code extensionRelated failures to investigate next
Root cause
- Overflow in loss computation (e.g., log(0) in cross-entropy)
- Mixed precision with incorrect loss scaling
- Division by very small number producing inf
The fix and how to prevent it
Evaluate Denpex on your own logs
Request a Scale evaluation code. A verified workplace organization activates 14 days with up to 50 diagnoses a day. Every account keeps its current diagnosis allowance and gets a verification path. No card or automatic subscription.
Don't just read the fix, diagnose your run
The encyclopedia tells you what went wrong. Denpex tells you what went wrong in YOUR training run. With your logs, your config, and your stack.